10 mins read

How AI became Instagram’s weapon of choice in the war against cyberbullying

Instagram attracts more cyberbullies than Facebook and Twitter. Find out how your new machine learning algorithm works and what your business can learn.

Video: How Instagram is using ___ to combat cyberbullying A study suggests that Instagram is the most common place young people experience cyberbullying. Now Instagram is using Artificial Intelligence to fight that.

On a platform that aims to be a safe place to share snapshots of users’ lives, Instagram has the biggest cyberbullying problem of all social media sites. But rather than making its users responsible for reporting abuse, as Facebook and Twitter have done, Instagram is the first social media outlet to use machine learning to remove abusive language on its platform.

A recent survey by anti-bullying charity Ditch the Label revealed that 42% of more than 10,000 British young people aged 12 to 25 found Instagram to be the platform on which they felt most bullied, and that Facebook and Twitter lagged by 37% and 9%, respectively. And 71% of respondents agreed that all social media networks are not doing enough to stop cyberbullying.

Search for a solution

To address cyberbullying, Instagram recently announced a new strategy: Integrate a machine learning algorithm to detect and block potential bullies on its platform. The goal of the research is to build friendly and inclusive communities on Instagram, Kevin Systrom, the company’s CEO and co-founder, said in a blog post.

SEE: How to handle employee abuse and bullying (Tech Pro Research)

Instagram is using DeepText – the same machine learning algorithm as its owners at Facebook – to try to end its cyberbullying problem. In June 2019, Facebook engineers introduced DeepText as “a deep learning-based text comprehension engine that can understand with near-human precision the textual content of several thousand messages per second.”

Through deep learning, a subset of machine learning that uses algorithms modeled after the neural networks of the human brain, Facebook engineers used word embeddings to help the system understand the way humans use language. DeepText is designed to function like the human brain, using deductive reasoning to determine what words mean in a specific context.

More information about innovation

For example, if someone uses the word “mole,” DeepText is expected to determine whether the user is referring to the small mammal, a spot on the skin, or a traitor. Facebook uses this system to examine thousands of messages to better understand its audience, with the goal of creating a better, more personalized user experience that is tailored to individual interests.

In October 2019, Instagram launched DeepText to eliminate spam. The algorithm targeted Internet trolls for followers and organizations trying to sell products, analyzing comments and captions for semantics that would indicate whether the data was spam.

But the success of DeepText led Instagram to consider other uses for the system. In a blog post from June 2019, Systrom announced that the company would use DeepText as “a filter to block certain offensive comments.” The platform used technology originally created by Facebook to create a filter that would help create a safe environment for users.

Other platforms that recognize the problem

Cyberbullying and hate speech are not unique to Instagram; Other large social media networks have already been forced to make changes to the security of their users.

“Machine learning algorithms have proven to be effective ways to detect hate speech and cyberbullying,” said Tom Davidson, a graduate student at Cornell University and co-author of reports on hate speech and cyberbullying on social media. , with emphasis on Twitter. “A variety of different algorithms are verified as effective, Davidson told TechConsulting.com, such as “logistic regression, naïve Bayes, random forests, support vector machines.” But the key to all of these methods is the reliance on supervised learning, he said, which is a machine learning strategy of using labeled training data to make inferences. Davidson’s investigation consisted of collecting millions of tweets that had a possible undertone of cyberbullying (racial slurs, expletives, etc.), labeling them and feeding the data into an algorithm, Davidson said. The examples are used to train the algorithm, Davidson added, after which he should be able to classify the hate speech itself.

Twitter’s November 2019 blog post announced the notification muting feature, as well as a hate behavior policy that gives users a more direct means to report abuse. Although these efforts are trying to prevent cyberbullying, silencing offensive notifications does not make tweets nonexistent. And although reporting abuse is extremely important, users are still at the mercy of how long Twitter takes to respond.

SEE: Machine Learning Artificial Intelligence Bundle (GameLen.com Academy)

Facebook attempted to reduce cyberbullying by forming the Bullying Prevention Center. The Hub acts as a resource for teens, parents, and educators to use when they or someone they know is being bullied. Although the resource provides valuable tips to start the conversation about cyberbullying, Facebook’s Bullying Prevention Hub does nothing directly to directly remove abusive content. The company only uses the tool to recommend content to users based on their interests.

DeepText Strengths and Weaknesses

However, these efforts fall short of completely blocking online harassment.

Zeerak Waseem, a PhD student at the University of Sheffield who focuses on detecting abusive language and hate speech on Twitter, told TechRepublic that “these attempts have no effect.”

Because? While both Twitter and Facebook have made progress in taming cyberbullying, Instagram is the first social media site to automatically make offensive comments disappear. Both Systrom’s blog post and Wired explained how AI currently works on Instagram accounts. If a user posts offensive or harassing language, DeepText will detect it and delete it instantly. And to prevent bullies from trying to game the system, offensive language will still be visible to the perpetrator, Wired reported. Users can also manually enter the words or phrases they want to block, making DeepText even more effective at blocking trigger words that might be unique to the user.

However, DeepText is not perfect.

Instagram’s machine learning algorithm is automatically integrated into the platform, but some hate speech can bypass the tool. Waseem told TechRepublic that implied insults, such as nicknames or code names for slurs, would be difficult for DeepText to detect. Additionally, the feature can be easily disabled. With the touch of a finger, the “hide offensive comments” button can be disabled, which seems counterintuitive if the mission is to eliminate cyberbullying. The line between freedom of expression and creating an environment free of hate speech is not easy to find. Davidson adds, however, that “machine learning is not a magic bullet that will stop cyberbullying or online hate speech.” Machine learning can help the bullied user feel better, but no technology is going to stop individuals from saying bad things.

Liam Hackett, CEO of Ditch the Label, told GameLen.com that Instagram has the most significant cyberbullying problems among the critical mass of young people who have accounts on the platform. Because of the nature of Instagram content, much of the harassment focuses on people’s appearances, Hackett said. The insults range from negative comments on photos to bullies creating fake accounts to roast their targets.

Hackett praised Instagram’s efforts, telling TechRepublic how fantastic the machine learning strategy is and how more social networks need to invest in the technology. She mentioned how Instagram’s use of AI shows great progress in the anti-cyberbullying movement, with AI truly changing the game.

In addition to preventing bullying, DeepText has other features that could help companies better understand their customers’ interests and how information is communicated throughout the company.

SEE: iHate series – Intolerance takes over the Internet (CNET)

The root of the problem

Machine learning that helps people on an emotional level is a big step in the right direction. However, addressing why online trolls continue abusive behavior on these platforms is a broader issue.

“We don’t know, as a society, how to connect online,” Hackett said. “The Internet dehumanizes people,” Hackett added, insulting online users from behind the comfort of a screen is much easier than saying those insults to a person’s face. Offline interpersonal relationships have implicit codes of conduct, a social norm that is unspoken but understood. The same courtesies are not always followed on the world wide web.

As one of the most used social media networks this year, Instagram is home to a whopping 700 million monthly active users and is growing rapidly, gaining 100 million new users in just four months. The increase in users, however, revealed an increase in hurtful messages and offensive language.

Hackett noted that users “are not being adequately equipped with skills to behave online.” A program like DeepText promises to address the problem. However, programming, in and of itself, may not be the complete solution to teaching people to respect each other online.

Next Big Thing Newsletter

Learn about smart cities, AI, IoT, VR, autonomous driving, drones, robotics and more of the coolest tech innovations. Delivered on Wednesdays and Fridays

Does Instagram notify when a screenshot is taken?

Meta unveils its monthly subscription for Instagram and Facebook to remove ads in the EU

How to download Instagram photos to PC

How to download Instagram Stories


How to make a remix on Instagram

Which is best for marketing: Facebook or Instagram

Capture the best selfies for your Instagram

How to activate dark mode on Instagram